The results chain and logframe: from inputs to impact

A results chain is the step-by-step logic of a project: inputs pay for activities, activities produce outputs, outputs lead to outcomes, and outcomes contribute to impact. A logframe, short for logical framework, puts that chain into a table and adds, for each level, the indicators you will measure, where the evidence will come from and the assumptions that must hold.

Both are standard tools for CSR teams and NGOs in India. Funders ask for them in proposals, monitoring reports are built on them, and a good impact assessment tests them. This guide explains each link in the chain, traces one invented project through it, and shows how to build a logframe, with a template you can copy.

If you are starting from scratch, build a theory of change first. It explains why the project should work; the results chain and logframe turn that reasoning into a plan you can deliver and measure.

The five links in a results chain

The OECD’s glossary of evaluation terms, a reference many funders and evaluators use, defines the links like this, in summary:

Two points in those definitions matter. Outputs are within the implementing team’s control, which is why they are easy to promise and count. Outcomes depend on choices people make and on other factors, which is why they are harder to deliver and the ones that really matter. And impact includes unintended and negative effects, which honest reports should look for too.

One project, traced through the chain

Notice where the chain can break. Some trainees drop out because the bus fare to the centre is more than they can manage. Some placements last two months because the job is far from home. A course can deliver every output and still fall short on outcomes, which is exactly what monitoring should catch.

Outputs or outcomes? The most common confusion

Most CSR reports are full of outputs presented as if they were results. A quick test: an output is something the project did or delivered; an outcome is a change in people’s lives, behaviour, skills in use or circumstances.

Count people once, say how you counted, and keep “people reached” separate from “people who benefited”. A trainee who attended one session was reached; one who now earns a living from the skill benefited.

The logframe matrix

A logframe has a row for each level of results and four columns: a summary of what is intended, the indicators, the means of verification, and the assumptions. In the classic format, the activities row lists the inputs and costs instead of indicators. Here is Kaushal Path’s.

Read it two ways. Upwards (the vertical logic): if the activities are carried out and their assumptions hold, the outputs follow; if the outputs are delivered and their assumptions hold, the outcome follows, and so on. Across (the horizontal logic): every level says how you will know it has happened.

Means of verification: where the evidence comes from

A means of verification names the source of each indicator: the register, survey, record or test the number will come from, who collects it and when. Vague entries such as “project reports” or “field visits” are a warning sign, because they don’t say how anyone could check the figure.

Good means of verification are:

• Specific: “attendance registers signed by the trainer each day”, not “records”.
• Affordable: a follow-up phone survey of every graduate may cost more than it is worth; a well-drawn sample may be enough.
• Independent where it matters: an employer’s confirmation or a school’s own records carry more weight than the project’s say-so.
• Planned from the start: a baseline taken before the work begins, and the same method used later. See baselines and evaluation.

Writing good assumptions

Assumptions in a logframe are conditions outside the project’s control that must hold for one level to lead to the next. Write them as positive conditions (“employers keep their placement commitments”), check that each is reasonably likely, and decide who will watch it.

If an assumption is both important and unlikely, don’t just record it: redesign. If trainees can’t afford the fare, add a travel allowance and turn the assumption into an activity. A logframe whose assumptions are all safe (“funding continues”) hasn’t been thought through.

Results chains in Indian practice

The government uses the same logic for its own spending. The Development Monitoring and Evaluation Office (DMEO) of NITI Aayog, set up in September 2015, finalises an Output-Outcome Monitoring Framework with ministries each year, before the Union Budget is laid before Parliament, to map the money for central schemes to their intended outputs and outcome targets. If your project works alongside a scheme, its framework can suggest indicators that officials already track.

For CSR, the law asks for less than you might expect. The annual action plan must set out a “monitoring and reporting mechanism” for each project (Rule 5(2) of the CSR Rules), and the board must satisfy itself that the money was used as approved, with the chief financial officer certifying it (Rule 4(5)). The statutory reports, the annual report on CSR and Form CSR-2, mainly record spending: apart from the impact assessment summaries required of the largest CSR spenders, they don’t ask for outcomes. Reporting outcomes is something good companies and NGOs choose to do, and our guide to monitoring CSR projects shows how.

A logframe template, and how to use it

A logframe is a management tool, not a contract carved in stone. Review it with partners at least once a year. When monitoring shows that an activity isn’t producing its output, or an output isn’t leading to its outcome, change the plan and agree the change with the funder. Our guide to choosing indicators helps you fill in the second column.